What's one decision this week where you trusted your judgment over the playbook? Where the "right answer" by any framework wasn't the one you went with? Walk me through what you weighed. (I'm publishing Sunday on the version of this that costs PMs interviews. If your story is good, I might pull from it — anonymously.) Subscribe to joshatlas.co/substack so the post lands in your inbox when it drops.

Most PMs are trained to seek clarity before acting — "Can you confirm the priority?" or "Should I proceed with X?" But strategic PMs lead with a hypothesis and check alignment: "I'm assuming X based on Y — does it make sense to proceed?" Share a specific moment where you asked for permission when you could have led with a hypothesis. What changed in how others perceived you, and what would you say differently today?

Hi everyone, I've lately started to use Replit and Claude Code for building clickable prototypes. I think it would be a lot of fun to get better at actually producing my own software. Is anyone else going down this route?

Hey everyone Michael here lovely to be here and learn from you all

The prompt lands — "Design a product to help people with home emergencies" — and the candidate immediately starts narrowing. Picks elderly. Or parents. Starts listing pain points. They skipped the most important step. Before you touch users, before you touch solutions, you need to show the interviewer that you understand why this problem exists right now. What trends created it. What makes it genuinely hard. For a home emergency prompt that means naming things like: smart home adoption is accelerating, the population is aging, mental health and personal safety concerns are rising. These aren't throat-clearing — they're the lens that makes every subsequent decision legible. Then you name what's hard about the prompt. Home emergencies aren't one thing. They're health crises, safety threats, personal security situations. The hardest part isn't building the solution — it's that different emergencies require completely different response mechanisms, and no single product has solved the multi-modal problem well. Now you pick your user segment. And your choice is defensible because it's grounded in the context you just built. That sequence — context, what's hard, then users, then solutions — is what product sense actually looks like under pressure. Most interview prep skips straight to step three. That's why most candidates sound competent but not exceptional. I've spent nearly 20 years in product at Meta, Google, PayPal, and Nest — and now I coach mid to senior PMs into the roles they're aiming for. This is the pattern I see most often.